[M3devel] AMD64_FREEBSD problems with current RC from m3 site

[Mika Nystrom]

Tony Hosking writes:
>Deadlock is not possible in this instance.  No other locks should be  
>held.  I notice that it used to be EAGAIN: does that imply we should  
>be retrying the lock here?
>
>Did you confirm linking with -lthr?

Yeah, the first (single ASSERT failure) was -lpthread.  The second
(double ASSERT failure) was -lthr.

     Mika